{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T06:12:18Z","timestamp":1788415938576,"version":"build-2803163510"},"reference-count":48,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T00:00:00Z","timestamp":1788393600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Big Data"],"abstract":"<jats:p>\n                    Financial correlation networks under infrastructure shocks undergo curvature-driven interference that standard estimators cannot detect. This study introduces Heat-Kernel Deep Instrumental Variables (HK-DeepIV), a geometric diagnostic and early-warning framework modeling shock propagation as heat diffusion on the Riemannian manifold of Johannesburg Stock Exchange (JSE) asset-return correlations. Three formal results underpin the architecture: a non-parametric identification result for the projection of the structural function onto the leading heat-kernel eigenfunction (\n                    <jats:italic>k<\/jats:italic>\n                    = 1; a scalar instrument can identify at most one linear combination of the eigenfunctions, and no claim is made beyond that projection); double robustness via Neyman orthogonality; and Corollary 3.2, showing that unit-level distinguishability collapses exponentially in diffusion time whenever Ollivier-Ricci curvature is positive; a purely geometric statement providing the basis for the fragility diagnostics developed here. Applied to 60 JSE tickers (2,832 trading days, 2015\u20132025) with Eskom load-shedding as the treatment (2022\u20132025), the trained model produces an ATE of +383 bp, reported as an overfitting artifact. Five independent baselines converge on smaller, mostly negative estimates. The most credibly conditioned conditional-association estimate is double machine learning [\u221220.75 bp, heteroskedasticity-and-autocorrelation-consistent (HAC)-corrected 95% CI [\u221251.62, 10.12] bp,\n                    <jats:italic>p<\/jats:italic>\n                    = 0.188], which is not significant at conventional levels; consistent with the instrument exogeneity caveat (5-day lagged return balance test,\n                    <jats:italic>p<\/jats:italic>\n                    = 0.007) and the descriptive framing of all estimates. The instrument [48-h-ahead Eskom stage forecast, Corr(\n                    <jats:italic>Z<\/jats:italic>\n                    <jats:sub>\n                      <jats:italic>t<\/jats:italic>\n                    <\/jats:sub>\n                    ,\n                    <jats:italic>D<\/jats:italic>\n                    <jats:sub>\n                      <jats:italic>t<\/jats:italic>\n                    <\/jats:sub>\n                    )\u22480.71, first-stage\n                    <jats:italic>F<\/jats:italic>\n                    = 47.3] is distinct from the treatment (realized Stage \u22652 binary), but exogeneity is not confirmed; all estimates are conditional associations. The Fiedler eigenvalue (mean 0.3827, minimum 0.1819) and mean Ollivier-Ricci curvature (0.5517, persistently positive) provide computable real-time fragility indicators. Across the four most severe load-shedding quarters, Fiedler and Ricci diagnostics offer comparable early-warning signals (mean lead-time difference \u22120.5 trading days); neither is systematically superior, but their combination is more informative than either alone. The framework contributes toward real-time spectral monitoring of infrastructure-driven systemic risk, supporting Uited Nations Sustainable Development Goal (UN SDG 9) in emerging markets.\n                  <\/jats:p>","DOI":"10.3389\/fdata.2026.1925931","type":"journal-article","created":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T05:13:46Z","timestamp":1788412426000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["HK-DeepIV: heat-kernel geometric diagnostics and early-warning signals for curvature-induced interference in financial correlation networks"],"prefix":"10.3389","volume":"9","author":[{"given":"Ntebogang Dinah","family":"Moroke","sequence":"first","affiliation":[{"name":"Department of Statistics and Operations Research, North-West University, Mafikeng Campus","place":["Potchefstroom, South Africa"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,3]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1977","DOI":"10.3982\/ECTA9623","article-title":"The network origins of aggregate fluctuations","volume":"80","author":"Acemoglu","year":"2012","journal-title":"Econometrica"},{"key":"B2","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1086\/262109","article-title":"Financial contagion","volume":"108","author":"Allen","year":"2000","journal-title":"J. Polit. Econ"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.1007\/978-4-431-55978-8","author":"Amari","year":"2016","journal-title":"Information Geometry and Its Applications"},{"key":"B4","doi-asserted-by":"publisher","first-page":"817","DOI":"10.2307\/2938229","article-title":"Heteroskedasticity and autocorrelation consistent covariance matrix estimation","volume":"59","author":"Andrews","year":"1991","journal-title":"Econometrica"},{"key":"B5","doi-asserted-by":"publisher","first-page":"444","DOI":"10.1080\/01621459.1996.10476902","article-title":"Identification of causal effects using instrumental variables","volume":"91","author":"Angrist","year":"1996","journal-title":"J. Am. Stat. Assoc"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1912","DOI":"10.1214\/16-AOAS1005","article-title":"Estimating average causal effects under general interference, with application to a social network experiment","volume":"11","author":"Aronow","year":"2017","journal-title":"Ann. Appl. Stat"},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2104.13478","article-title":"Geometric deep learning: grids, groups, graphs, geodesics, and gauges","author":"Bronstein","year":"2021","journal-title":"arXiv Preprint arXiv"},{"key":"B8","doi-asserted-by":"publisher","first-page":"331","DOI":"10.1007\/s10208-006-0196-8","article-title":"Optimal rates for the regularized least-squares algorithm","volume":"7","author":"Caponnetto","year":"2007","journal-title":"Found. Comput. Math"},{"key":"B9","doi-asserted-by":"publisher","first-page":"C1","DOI":"10.1111\/ectj.12097","article-title":"Double\/debiased machine learning for treatment and structural parameters","volume":"21","author":"Chernozhukov","year":"2018","journal-title":"Econometr. J"},{"key":"B10","doi-asserted-by":"publisher","first-page":"e41010","DOI":"10.1371\/journal.pone.0041010","article-title":"Methods for detecting early warnings of critical transitions in time series illustrated using simulated ecological data","volume":"7","author":"Dakos","year":"2012","journal-title":"PLoS ONE"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1606.09375","article-title":"Convolutional neural networks on graphs with fast localized spectral filtering","author":"Defferrard","year":"2016","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"B12","doi-asserted-by":"publisher","first-page":"2153","DOI":"10.1007\/11503415_22","article-title":"On the Nystr\u00f6m method for approximating a Gram matrix for improved kernel-based learning","volume":"6","author":"Drineas","year":"2005","journal-title":"J. Mach. Learn. Res"},{"key":"B13","first-page":"78","article-title":"Impact of loadshedding in South Africa: A CGE analysis","volume":"10","author":"Erero","year":"2023","journal-title":"J. Econ. Polit. Econ."},{"key":"B14","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3309547","article-title":"Temporal relational ranking for stock prediction","volume":"37","author":"Feng","year":"2019","journal-title":"ACM Trans. Inf. Syst."},{"key":"B15","doi-asserted-by":"publisher","first-page":"901","DOI":"10.1080\/01621459.2020.1768100","article-title":"Identification and estimation of treatment and interference effects in observational studies on networks","volume":"116","author":"Forastiere","year":"2021","journal-title":"J. Am. Stat. Assoc"},{"key":"B16","doi-asserted-by":"publisher","first-page":"779","DOI":"10.1257\/jel.20151228","article-title":"Contagion in financial networks","volume":"54","author":"Glasserman","year":"2016","journal-title":"J. Econ. Lit"},{"key":"B17","doi-asserted-by":"publisher","first-page":"297","DOI":"10.1007\/s40471-016-0086-4","article-title":"Dependent happenings: a recent methodological review","volume":"3","author":"Halloran","year":"2016","journal-title":"Curr. Epidemiol. Rep"},{"key":"B18","first-page":"1414","article-title":"\u201cDeep IV: A flexible approach for counterfactual prediction,\u201d","author":"Hartford","year":"2017","journal-title":"International Conference on Machine Learning (PMLR)"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1412.6980","article-title":"\u201cAdam: a method for stochastic optimization,\u201d","author":"Kingma","year":"2015","journal-title":"Proceedings of the 3rd international conference on learning representations (ICLR)"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1609.02907","article-title":"\u201cSemi-supervised classification with graph convolutional networks,\u201d","author":"Kipf","year":"2017","journal-title":"Proceedings of the 5th international conference on learning representations (ICLR)"},{"key":"B21","doi-asserted-by":"crossref","DOI":"10.1090\/mbk\/107","volume-title":"Markov Chains and Mixing Times, 2nd Edn","author":"Levin","year":"2017"},{"key":"B22","doi-asserted-by":"publisher","first-page":"605","DOI":"10.2748\/tmj\/1325886283","article-title":"Ricci curvature of graphs","volume":"63","author":"Lin","year":"2011","journal-title":"Tohoku Math. J"},{"key":"B23","doi-asserted-by":"publisher","first-page":"S1","DOI":"10.1111\/j.1368-423X.2012.00368.x","article-title":"Identification of treatment response with social interactions","volume":"16","author":"Manski","year":"2013","journal-title":"Econometr. J"},{"key":"B24","doi-asserted-by":"publisher","first-page":"193","DOI":"10.1007\/s100510050929","article-title":"Hierarchical structure in financial markets","volume":"11","author":"Mantegna","year":"1999","journal-title":"Eur. Phys. J. B"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.21887713","author":"Moroke","year":"2026","journal-title":"IST02-IST03 JSE-Eskom Infrastructure-Coupled Financial Network Dataset (Version V4.0) [Dataset]"},{"key":"B26","first-page":"2710","article-title":"\u201cDual instrumental variable regression,\u201d","author":"Muandet","year":"2020","journal-title":"Advances in neural information processing systems (NeurIPS)"},{"key":"B27","doi-asserted-by":"publisher","first-page":"7","DOI":"10.1080\/0376835X.2024.2408364","article-title":"The impact of load shedding on the South Africa economy","volume":"58","author":"Naidoo","year":"2023","journal-title":"J. Public Adm."},{"key":"B28","doi-asserted-by":"publisher","first-page":"1565","DOI":"10.1111\/1468-0262.00459","article-title":"Instrumental variable estimation of nonparametric models","volume":"71","author":"Newey","year":"2003","journal-title":"Econometrica"},{"key":"B29","first-page":"25956","article-title":"\u201cRevisiting over-smoothing and over-squashing using Ollivier-Ricci curvature,\u201d","volume":"202","author":"Nguyen","year":"2023","journal-title":"Proceedings of the 40th international conference on machine learning (ICML)"},{"key":"B30","doi-asserted-by":"publisher","first-page":"299","DOI":"10.1093\/biomet\/asaa076","article-title":"Quasi-oracle estimation of heterogeneous treatment effects","volume":"108","author":"Nie","year":"2021","journal-title":"Biometrika"},{"key":"B31","doi-asserted-by":"publisher","first-page":"105729","DOI":"10.1016\/j.frl.2024.105729","article-title":"Blackouts and stock markets: evidence from load-shedding in South Africa","volume":"67","author":"Obalade","year":"2024","journal-title":"Fin. Res. Lett"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.2022.2131557","article-title":"Causal inference for social network data","author":"Ogburn","year":"2024","journal-title":"J. Am. Stat. Assoc"},{"key":"B33","doi-asserted-by":"publisher","first-page":"810","DOI":"10.1016\/j.jfa.2008.11.001","article-title":"Ricci curvature of Markov chains on metric spaces","volume":"256","author":"Ollivier","year":"2009","journal-title":"J. Funct. Anal"},{"key":"B34","first-page":"1595","article-title":"\u201cDYNOTEARS: structure learning from time-series data,\u201d","author":"Pamfil","year":"2020","journal-title":"International Conference on Artificial Intelligence and Statistics (PMLR)"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511803161","author":"Pearl","year":"2009","journal-title":"Causality: Models, Reasoning, and Inference, 2nd Edn"},{"key":"B36","volume-title":"Elements of Causal Inference","author":"Peters","year":"2017"},{"key":"B37","doi-asserted-by":"publisher","first-page":"688","DOI":"10.1037\/h0037350","article-title":"Estimating causal effects of treatments in randomized and nonrandomized studies","volume":"66","author":"Rubin","year":"1974","journal-title":"J. Educ. Psychol"},{"key":"B38","doi-asserted-by":"publisher","first-page":"e1501495","DOI":"10.1126\/sciadv.1501495","article-title":"Ricci curvature: an economic indicator for market fragility and systemic risk","volume":"2","author":"Sandhu","year":"2016","journal-title":"Sci. Adv"},{"key":"B39","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1038\/nature08227","article-title":"Early-warning signals for critical transitions","volume":"461","author":"Scheffer","year":"2009","journal-title":"Nature"},{"key":"B40","doi-asserted-by":"publisher","first-page":"765","DOI":"10.1145\/3501714.3501755","article-title":"\u201cCausality for machine learning,\u201d","author":"Sch\u00f6lkopf","year":"2022","journal-title":"Probabilistic and Causal Inference: The Works of Judea Pearl"},{"key":"B41","doi-asserted-by":"publisher","first-page":"1040","DOI":"10.1214\/aos\/1176345969","article-title":"Optimal global rates of convergence for nonparametric regression","volume":"10","author":"Stone","year":"1982","journal-title":"Ann. Stat"},{"key":"B42","doi-asserted-by":"crossref","DOI":"10.1145\/3705328.3759312","article-title":"\u201cRicciFlowRec: a geometric root cause recommender using Ricci curvature on financial graphs,\u201d","volume-title":"Proceedings of the 19th ACM conference on recommender systems","author":"Sun","year":"2025"},{"key":"B43","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2111.14522","article-title":"\u201cUnderstanding over-squashing and bottlenecks on graphs via curvature,\u201d","author":"Topping","year":"2022","journal-title":"Proceedings of the 10th international conference on learning representations (ICLR)"},{"key":"B44","doi-asserted-by":"publisher","first-page":"659","DOI":"10.1016\/j.jempfin.2010.04.008","article-title":"A network perspective of the stock market","volume":"17","author":"Tse","year":"2010","journal-title":"J. Empir. Fin"},{"key":"B45","doi-asserted-by":"publisher","first-page":"1579332","DOI":"10.3389\/fdata.2025.1579332","article-title":"Urban mobility and crime: causal inference using street closures as an instrumental variable","volume":"8","author":"Vachuska","year":"2025","journal-title":"Front. Big Data"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1710.10903","article-title":"\u201cGraph attention networks,\u201d","author":"Velickovic","year":"2018","journal-title":"Proceedings of the 6th international conference on learning representations (ICLR)"},{"key":"B47","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.07154","article-title":"\u201cLearning deep features in instrumental variable regression,\u201d","author":"Xu","year":"2021","journal-title":"Proceedings of the 9th international conference on learning representations (ICLR)"},{"key":"B48","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2506.02790","article-title":"Orthogonality-constrained deep instrumental variable model for causal effect estimation","author":"Yao","year":"2025","journal-title":"arXiv Preprint arXiv"}],"container-title":["Frontiers in Big Data"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2026.1925931\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T05:13:49Z","timestamp":1788412429000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2026.1925931\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,3]]},"references-count":48,"alternative-id":["10.3389\/fdata.2026.1925931"],"URL":"https:\/\/doi.org\/10.3389\/fdata.2026.1925931","relation":{},"ISSN":["2624-909X"],"issn-type":[{"value":"2624-909X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,3]]},"article-number":"1925931"}}